What an inclusion list is for and what it costs
An exclusion list removes named domains, apps or placements from an otherwise open buy, which makes it reactive: a site that is not yet on it is still bought. An inclusion list works the other way round. The demand-side platform only bids where the list allows, so a newly registered made-for-advertisingA site built to sell ad impressions rather than serve readers: thin or scraped content, a high ratio of ads to content and traffic bought rather than earned. domain is kept out without anyone having to notice it first. That is its value, and the fundamentals hub shows where in the journey of an impression it acts: before the bid, alongside the other targeting filters.
Its cost is the other side of the same mechanism. Every domain left off is inventory the buy can no longer reach, so ReachUnique people who saw an ad. falls and the price per thousand rises as the platform competes for a smaller pool. The brand safety guide sets out that trade and the comparison to make once controls are in place: cost per impression that met the standard, or cost per outcome, rather than raw cost per thousand.
| Approach | What the buy can reach | Main risk | Maintenance |
|---|---|---|---|
| Open buy with exclusion lists | Everything except what has been named | New and unfamiliar domains are bought until someone notices them | Grows as problems appear; rarely shrinks |
| Broad inclusion list | A large named set chosen from the audience's reading | Reach shortfall if the set is too narrow for the budget | Additions as the audience's reading changes |
| Short inclusion list | A small set of trusted publishers | Under-delivery and high prices | Close watch on delivery and price per domain |
| Deals with chosen publishers | The inventory each deal carries | Price premium; pages within a publisher still vary | Deal terms and delivery, as the deal types guide explains |
The list does not replace the other controls. Content labels, sensitive category exclusions and placement exclusionsLists of websites, apps, video channels or app categories where an advertiser's ads must not appear. still decide which pages inside a listed domain are acceptable, and the written tolerance those settings come from belongs to the brand suitability policy. Made-for-advertising inventory, how to recognise it and the vendor segments that classify it stay in the brand safety guide.
Before you start
- The objective and the plan: what the buy is for, its budget and flight, and whether it is measured on reach or on an outcome on the advertiser's site.
- The written suitability policy, or at least the brand owner's agreed stance on news, user-generated pages and sensitive topics. Without it, the judgement step below becomes one person's taste.
- Past placement or site reports at domain and app level with spend and outcomes, from DV360's offline reports as the DV360 reporting guide describes them.
- Audience research: what the customers read, watch and use, from surveys, sales conversations, site referral data or the planner's own work.
- Access to the DV360 advertiser, and to the partner if the list is to be shared across advertisers.
- A spreadsheet for the candidate register, with a column for the reason each entry is there.
The steps
Decide from the objective what the list must achieve
Write one sentence that says what the list is for. A reach campaign needs a list broad enough to deliver its budget among the right people; a campaign for a brand with real reputational exposure, or one measured on outcomes, can accept a smaller pool at a higher price because the removed inventory would not have produced the outcome anyway. The sentence decides how wide to cast the net and how hard to cut later, and it stops the list from turning into a catalogue of every respectable site the team can name. Note which line items the list will govern: prospecting, remarketing and deal line items often need different breadth.
Gather candidates from where the audience reads
Start from the people, not the inventory. List the publishers, sections and apps the audience is known to use: the national and local news they read, the specialist titles of their interests, the apps they open daily. Add categories of publisher rather than single names where the audience is spread across many small titles, and then find the titles within each category. Vendor and industry lists are acceptable as a source of further candidates, but they enter the register on the same terms as everything else. Keep news in: the brand safety guide warns that treating news as unsafe pushes budget towards inventory that reports on nothing.
Add publishers that past placement reports show earned outcomes
Pull the site and app report for previous buys and sort it by spend, then read the outcome columns beside it: engaged visits, conversions or whatever the measurement plan counts. Domains that took money and returned outcomes are strong candidates even if nobody would have thought of them; domains that took money and returned nothing are evidence for the exclusion list, not this one. Read by spend rather than impressions, as the brand safety guide advises, so a long tail of cheap unfamiliar domains cannot hide where the budget went. If the business also buys display in Google Ads, the guide to low-quality display placements shows how to read the same signals there. Where that report already puts most of the spend on domains nobody chose, the diagnosis of programmatic spend going to low-quality sites names the other doors, the goal, the verification, the supply path and the inherited settings, that this list alone does not close.
Check each domain's authorised sellers and supply path
For each candidate, check who is allowed to sell it. A publisher's ads.txt file lists the companies authorised to sell its inventory and whether each is a direct seller or a reseller, and the programmatic ecosystem guide explains how
sellers.jsonand the SupplyChain object in the bid request show who stands behind each seller and every hop an impression took; the real-time bidding guide shows where that object sits in the request. A domain with no ads.txtA file on a publisher's site listing the companies authorised to sell its ad inventory. file, or one sold only through resellers, deserves a closer look before it earns a place. The DV360 inventory and deals guide describes authorised-seller targeting, which uses ads.txt and app-ads.txt to limit buying to declared sellers, with options the Help Centre explains; decide which option the list's line items will use and record it.Judge each candidate against the suitability policy
Open each candidate and read it as the brand would. Does its content sit within the policy's stance on news, user-generated pages and sensitive topics? Is it a real publication with readers, or does it show the signs the brand safety guide lists for sites built to sell impressions, such as stacked and auto-refreshing units that score well on ViewabilityShare of measured ad impressions that had a real chance to be seen. while serving nobody who chose to be there? Record the judgement as a short reason in the register, not a tick. Where the policy is silent on something the candidate raises, note the question and take it to the policy's owner rather than deciding it in the spreadsheet; the answer belongs in the policy, where the next list builder will find it.
Score the candidates and cut to a list
Score the surviving candidates with a weighted scoring matrix: agree the criteria and their weights before anyone scores, then score each candidate with evidence in the cell. Useful criteria are audience fit, the quality of the content, fit with the suitability policy, the seller path and past outcomes. Treat a failed seller check or a clear breach of the policy as a gate rather than a low score, so no candidate can buy its way back with points elsewhere. Then cut from the bottom of the ranking, checking that the order survives a modest change of weights; a domain whose place depends on one weight is a judgement to discuss, not a number to trust.
Estimate the reach the list leaves before launch
Before the list goes live, estimate what it will deliver. Use DV360's inventory availability forecasts where they are offered for the line item's settings, and the past spend on listed domains from the placement reports. The impression calculator turns a budget and an expected price per thousand into impressions, so the plan can see whether the listed inventory can absorb the budget at a price it can bear. If it cannot, widen the list from the next candidates in the ranking or lower the budget on the governed line items, and write the expected reach and price into the plan so the result can be judged against them later. The academy checkpoint on suitability policies, inclusion lists and incidents tests sizing the reach a list leaves and recording a reason for every entry.
Build it as a shared channel and assign it
In DV360, the inventory guide describes channels as reusable lists of URLs and apps that can be included or excluded on line items, and the account structure guide notes that some partners keep shared channels at partner level so every advertiser can use them. Build the list once as a channel, at the advertiser for a single brand or at the partner where several advertisers share it, and include it on the line items the first step named. A list pasted into each line item separately drifts as soon as someone edits one copy. Check the build against the plan with the DV360 campaign launch checklist, which compares site and app lists with the inclusion and exclusion plan before the flight starts.
Record why each entry is on the list and review it from placement reports by spend
Keep the register beside the channel: each domain or app, the reason it was added, its score, the date and who decided. Give the list a version and note every change against it. Then review it on the rhythm the programmatic campaign management playbook sets for placement and site reports: read the report by spend, look at what each listed domain delivered against the plan's measure, retire entries that take money without results, and add the next candidates when the list cannot absorb the budget. Each review ends with a new version and a line in the change log, so anyone can see what the list held on any given day.
A hearing aid retailer builds its first list (illustrative)
Vexhollum Hearing Aids replaces an open buy with an inclusion list
Vexhollum Hearing Aids, an invented retailer with a website and 35 fitting centres, has run open-auction display for three months with sensitive category exclusions only. The site report shows £18,000 spent across 2,900 domains, with 47 per cent of spend on domains nobody on the team recognised and almost no hearing test bookings from them.
The marketing manager and the agency trader build a list over two working sessions.
| Stage | What happened |
|---|---|
| Objective | Hearing test bookings among people over 55 in the fitting centres' regions; reach matters less than bookings |
| Candidates | 310 gathered from health, retirement, local news and puzzle publishers, plus 60 domains that had produced bookings in the report |
| Seller check | 42 removed: no ads.txt file or sold only through resellers |
| Suitability | 25 removed against the policy, mostly user-generated pages; hard news kept |
| Scoring | Weights agreed first: audience fit 30, content 20, suitability 20, seller path 15, past outcomes 15; list cut to 180 |
| Reach estimate | Past spend on the 180 domains plus the forecast suggested about £6,000 a month at a £5.50 CPM, roughly 1.1 million impressions, against a £6,000 monthly budget |
| Build | One shared channel included on both prospecting line items |
| Review | Site report by spend on the first working day of each month, with a versioned change log |
In the first month the list delivers £5,700 of the budget at a £5.80 CPM. Hearing test bookings from display rise from 9 to 31. The monthly review removes 12 domains that took spend without a booking and adds 15 from the next candidates in the ranking.
Common mistakes
- Adopting a borrowed list wholesale. Nobody can say why each entry is there, so nobody can review it.
- Building the list from inventory rather than the audience. The result is every respectable site the team can name, not the ones its customers read.
- Leaving news off. It removes quality journalism and pushes budget to sites that report on nothing.
- Skipping the seller check. A trusted name sold through an unexpected reseller path is not the impression the list meant to buy.
- Scoring without agreeing weights first. Weights chosen after the scores tend to confirm the favourites.
- Launching without a reach estimate. The shortfall shows up mid-flight and is blamed on the market.
- Pasting the list into each line item. Copies drift; build one shared channel and assign it.
- Treating the list as finished. A list that only grows, and never records why, ends up as wide as the open web and as hard to review.
